SNES-Mini RGB Measurements

Great job on that The pictures comparison looks good. One may think of adding the parallel pull-down resistors on the small PCB - should be possible

However...

Be carefull - first, the values one the 1Chip board are different on PAL and NTSC boards to my knowledge. I may verify that at home Further, there might be a pin at the S-CPUN to control the Vpp-output for the RGB signals. I don't know the pinout, but on the S-PPU2 pin94 can be used to change the supply voltage for RGB.

Side note to verify Bad_Ads scetch: Here is a stripped PCB where you can see how the signal wires go

 

The values are the same on my Super Famicom 1CHIP as they are on my PAL 1CHIP. It would be interesting if there are any boards out there with different values.

I wonder if the possible pin at the S-CPUN to control the Vpp-output for the RGB signals that you described could have something to do with the difference in value of R6, R7, and R8 between the 1CHIP and the mini...just a thought.

 

Then it might be just the different values between 1Chip and Mini in my mind. However - I can look for the values on my 1Chip-01, 1Chip-02 (both PAL), 1Chip-01 (SFC). Will do it soon ^^

 

Nice pic borti. Yeah, I can't find any info about the S-CPUN chip either. Would be nice to know if the resistors between models are different too.

 

So, I ended up reinstalling the 150 ohm resistors, and went the way kel suggested. This is easier, since it doesn't involve replacing SMD components.

Used 1.1KOhm 1% tolerance resistors. I bought mine from Digi-Key (Part # S1.1KCACT-ND) .

EDIT: For more exact colors, you can use 1.2kOhm 1% tolerance resistors. Digikey does not have 1% tolerance at 1.2k, but Mouser does (Part # 660-MF1/4DCT52R1201F).

Connected one side to the vias connected to the RGB lines coming off the S-CPUN chip. Tied the other sides together, and soldered them to GND. I used the large spot nearby, that might have an electrolytic cap on some models (maybe PAL?). It's labeled on the front side of the board. This spot took a while to heat up. There are a few other spots nearby you can also use.



I checked the results at the Multi-Out. I get the same results (as expected) as I did when replacing the 150 ohms with 133 ohms. Much easier!

 

Nice pics Ultron. They reminded me that I really need to pull my finger out and give this a go.

 

OK, last measurements. I installed the THS7314 board sold by retrorgb on his site: http://www.retrorgb.com/store.html. This board has the THS7314 amp, along with the 75 ohm resistors on the output. 220uF caps are still needed in the cable.

The 1.1k ohm resistors were still used. I connected wires on the via side of the resistors, and connected them to the RGB amp board.



Measured the R, G, and B (left to right) off the Multi-Out.




2 things of note, the offset voltage (~240mV) is much lower than when using the outputs of the S-RGB (~2V offset). Not a problem, just something to note. The second thing is the overshoot when the signal changes. The THS7314 does have some overshoot on as the data sheet shows. It's been a while since I've messed with overshoot (since college!). It could be the chip, the location and thickness of the board traces, or it might need another decoupling cap on the supply (data sheet for THS7314 suggests 0.1uF and a 100uF parallel across the supply and GND pin, this board only has one, I think a .1uF?). I didn't see an effect on the image, but didn't check it for long.


Anyways, measurements are done. I can write instructions up for a proper mod now.

 

Awesome job!

 

Just looking at the PCB it doesn't seem that the decoupling cap would be effective. Adding 100 uF would be redundant since low frequency decoupling should already be provided by the SNES.

"220uF caps are still needed in the cable."

The THS can easily handle the offset current so technically there is no need for output ac-coupling, though surely it's best to keep cables as they are. There is a benefit to dc-coupling though--no field tilt.

 

Does anyone get slight diagonal interference in RGB from their SNES-mini / SFC Jr? I did until I found a jumper to bridge that totally eliminates it (eliminates chroma in composite too, though...)

Anyway, here's a pic of the location to bridge (C48):


You can experiment with putting in a cap, varicap, switch, whatever. I just did a simple solder bridge.

The change in quality is slight but noticeable in flat blue or green areas of colour:

 

Hmm, I have to check this. How did you figure this out?

 

I've never seen this on any of my RGB monitors or on an XRGB Mini, but I've seen this with capture cards. Is anyone else seeing it too?

 

Old-fashioned poking at random solder pads with a grounded wire.

It did seem like a crosstalk/grounding problem anyway. I use an X-RGB 2+ for RGB and it's sensitive to noise on the sync/compsync line, so that might partly be at fault.

 

I have seen those lines on quite a few SNES. But never on the very early SFC's with the sepearate audio module and never on 1Chip models, just the revisions in between had the diagonal lines for me.

 

My 1CHIP-01 does this too. Especially noticeable on an all green screen.

Any idea where we would find the equivalent soldering pad on a regular SNES?

Also, what's the correct way to wire up C.sync on an EU console? On my old SNES, I modified years ago, I just hooked up the C.sync pin directly to the AV-out. This works, but I assume it's not the correct way to do this

 

Have you tried just using a cable that uses composite sync instead of composite video for sync? I'm pretty sure what you're doing is just eliminating the chroma signal from composite video via a shunt, which is what causes the interference. If you just use composite sync instead, which doesn't have chroma data, you should eliminate the interference.

 

If I remember right, I think there is no easy way to use CSYNC on a PAL console. Pin 3 (for CSYNC on NTSC) Multi-Out on PAL is connected to +12V to tell the TV to use RGB I believe. You might be able to use a sync stripper LM1881 on the Composite Video pin 9 (or pin 20 on SCART side) and put the sync on that output pin, but you would lose the ability to use composite video. No idea if this would work, maybe a PAL guy can help you out more.

 

Luma carries a sync signal too and can be used instead of csync.

 

I tend to cut composite video and put luma or csync on the pin - this way all rgb cables still work without needed to modify the console end.

 

I have a cable that uses composite sync instead of composite video and this is indeed true.